Methodology

Divisions voted

"Voted in n of m divisions" counts the divisions in the member's own chamber held while they sat (m) and the divisions where their name appears in the ayes or noes (n). A member sworn in at a by-election, or one who has since left, is measured against their own time in the seat, not the whole parliament.

Absence is not abstention. The official record does not distinguish pairing arrangements, approved leave, ministerial or committee duties, or illness from any other reason for not voting. A low attendance figure is not, by itself, evidence of anything. This caveat is shown wherever the figure appears.

Votes against party majority ("crossed")

For each division, each parliamentary group's majority position is the side (aye or no) with more of that group's votes. A member's vote is marked "crossed" when it lands on the other side. Groups that split evenly have no majority position, so no vote in that division is marked. Independents have no group majority and are never marked.

Chamber composition

Seat counts group members by parliamentary group as recorded on Wikidata, which follows the official record: Coalition members are shown as one group. Counts reflect current sitting members, not the election result.

Election percentages

First preference and two-candidate-preferred percentages are computed from AEC final totals per electorate. Candidate shares are of formal votes, the base the AEC's own percentages and swings use. Informal ballots are shown after the candidates as a share of all ballots cast, beside the AEC's swing in informality. Swing figures are the AEC's own published swings.

Senate results

A senator's election history gives their group's first preferences as a share of the state's formal vote: above-the-line votes for the group plus every group candidate's own first preferences, over all formal votes in the state. Most Senate votes are cast for a group, so a candidate's personal count would understate the support they were elected on. Places elected are the AEC's order of election.

Results at each election

An electorate's list of past results names the winner of every contest for a division of that name since 2004, with their share of the two-candidate-preferred count where the AEC published one. Each contest was held on the boundaries then in force, so the list follows a name, not an unchanged area. Election history on a profile is matched by name within the member's state; a candidate on the roll under a longer first name ("Raffaele" for "Raff") counts only where exactly one member fits.

Expenses

Parliamentary expenses are IPEA's published transactions, summed per parliamentarian and per IPEA category for each of the latest four quarters and rounded to the dollar. Travel distances, office locations and roles differ, so totals are not compared or ranked.

Funding disclosures

Party returns are listed as lodged with the AEC, one row per return, never added together: a party's branches lodge separately and money passed between them counts in each. Donations are the gifts a party's returns list as donations received, summed by donor across the party's returns for the latest financial year. A member's returns are matched by name with titles removed; candidate returns as election history is.

Maps

Electorate maps draw the ABS's approximation of each division, built from mesh blocks, simplified to what the map can show. The AEC's own maps show the legal boundary.

AI summaries and notes

Machine-written text appears in four places, always labelled "AI-generated".

Bill context. Each bill's page explains the official summary in plain terms. General knowledge of Australian parliamentary practice is used to explain the mechanism (what an appropriation bill is, for example); anything specific to the bill comes from its official summary only. A bill with no official summary yet gets no note; one is written once APH publishes a summary.

Division context. They Vote For You provides written context for most bill votes; for procedural motions their context field is a raw Hansard excerpt. For those, a one-to-two sentence note explains what the bill, amendment or motion was about and what question was being decided, grounded on that excerpt and the official bill summaries. It never restates the result, which the page already shows, and links to the full record.

Series of divisions. When a chamber divides on one matter several times in a sitting day, the home page and the divisions index fold those divisions into one entry; the index's counts are still of divisions, not entries. On the home page each vote in a series is labelled with the first sentence of its context: They Vote For You's where volunteers have written it, otherwise the machine-written division note, marked AI. The sequence of amendments put and lost before the question itself then reads as one story. The divisions index names each vote by its official stage, or by its division number, and describes none.

Voting record in brief. Each member's page carries a short note describing patterns their voting table cannot show at a glance. The arithmetic is done first: every division the member voted in is read for the bill or kind of motion and the stage, and their ayes and noes are counted per bill and per kind. The generator is handed those counts and the divisions where they voted against their own party grouping, and nothing else, so it can phrase the sums but not make them; a note that uses a number the counts do not contain is discarded. It is instructed to describe, never evaluate: no praise, no criticism, no motives, no ideology. A note is rewritten once the member has voted in ten more divisions since it was written, or at once when they vote against their party grouping.

AI text is never part of the record. If a summary or note misstates the record, request a correction and it will be regenerated or removed.

What this site never does

No scoring, no ranking, no summarising of speeches, no inference of positions from votes. Where They Vote For You provides a plain-English description of a motion, it is shown with explicit attribution to them.